Towards a Minimal Abductive Reasoning Layer for SASIS

Aus SDQ-Wiki
Ausschreibung (Liste aller Ausschreibungen)
Typ Bachelorarbeit
Aushang Bachelor Thesis Towards a Minimal Abductive Reasoning Layer for SASIS.pdf
Betreuer Wenden Sie sich bei Interesse oder Fragen bitte an:

Nicolas Schuler (E-Mail: nicolas.schuler@kit.edu, Telefon: +49-721-608-46537), Vincenzo Scotti (E-Mail: vincenzo.scotti@kit.edu)

Motivation

Abductive reasoning – inferring the most plausible cause from observed effects – is a powerful paradigm for tasks like fault diagnosis and explanation generation. However, classical logic programming tools like Prolog are cumbersome to integrate into modern software systems, and expertise in these tools is increasingly rare. This thesis explores whether a minimal, modular abductive reasoning layer can be designed that integrates easily with existing software engineering frameworks and methodologies. The goal is to identify the essential formal components required for practical abduction and to evaluate the approach on concrete use cases in self-adaptive systems.

Tasks

  • Survey existing approaches to abductive reasoning (logic programming, ASP, SMT-based encodings) and analyze their suitability for lightweight integration.
  • Design and implement a minimal abductive reasoning layer that can be plugged into existing frameworks with low overhead.
  • Evaluate the approach on two distinct use cases (e.g., fault localization, configuration diagnosis, or adaptation explanation) to assess generality and practicality.

Tools / Technology

Python, Logical and Probabilistic Programming

Note: Thesis offered in German or English